| Abstrakt: |
Sentiment analysis is a method of calculating polarization of statements such as positive, negative and neutral comments of the posts, micro blogs, opinions and feedbacks which are published in any social media sites. In this world of application sites, the feelings of tweets or retweets published on Twitter are classified in this study by examining specific catchwords in tweets & finding the propagation as positive and vague. Feature selection is made to access the tweets. For the selection of excellent features, the class algorithms are used to teach and check the capabilities of words, as well as to decide the sentiment polarity of each tweet. The three machine learning classifiers namely Random Forest, Naive Bayes Classifier, and Support Vector machine are compared with accuracy, precision, and time. In this study, the results obtained were in favour of Support Vector Machine with accuracy (0.8) and F1-score (0.8). [ABSTRACT FROM AUTHOR] |
